{"vulnerability": "CVE-2025-38073", "sightings": [{"uuid": "b3f4aabb-08bd-4e11-9df5-40ab11dd27c9", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "2a075640-a300-48a4-bb44-bc6130783b9b", "vulnerability": "CVE-2025-38073", "type": "seen", "source": "https://bsky.app/profile/cve.skyfleet.blue/post/3lruvyydxlo2p", "content": "", "creation_timestamp": "2025-06-18T11:33:06.093079Z"}, {"uuid": "28158ad5-4995-4ab5-9d13-0b26d10c9a6a", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "2a075640-a300-48a4-bb44-bc6130783b9b", "vulnerability": "CVE-2025-38073", "type": "seen", "source": "https://vulnerability.circl.lu/bundle/816dcc8e-f25a-4895-9b59-1bbd9caeccb8", "content": "", "creation_timestamp": "2025-12-03T14:14:49.267740Z"}, {"uuid": "70b32202-99bf-4ff1-9ed4-d575ee82eb53", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "2a075640-a300-48a4-bb44-bc6130783b9b", "vulnerability": "CVE-2025-38073", "type": "published-proof-of-concept", "source": "https://t.me/DarkWebInformer_CVEAlerts/18708", "content": "\ud83d\udd17 DarkWebInformer.com - Cyber Threat Intelligence\n\ud83d\udccc CVE ID: CVE-2025-38073\n\ud83d\udd25 CVSS Score: N/A\n\ud83d\udd39 Description: In the Linux kernel, the following vulnerability has been resolved:\n\nblock: fix race between set_blocksize and read paths\n\nWith the new large sector size support, it's now the case that\nset_blocksize can change i_blksize and the folio order in a manner that\nconflicts with a concurrent reader and causes a kernel crash.\n\nSpecifically, let's say that udev-worker calls libblkid to detect the\nlabels on a block device.  The read call can create an order-0 folio to\nread the first 4096 bytes from the disk.  But then udev is preempted.\n\nNext, someone tries to mount an 8k-sectorsize filesystem from the same\nblock device.  The filesystem calls set_blksize, which sets i_blksize to\n8192 and the minimum folio order to 1.\n\nNow udev resumes, still holding the order-0 folio it allocated.  It then\ntries to schedule a read bio and do_mpage_readahead tries to create\nbufferheads for the folio.  Unfortunately, blocks_per_folio == 0 because\nthe page size is 4096 but the blocksize is 8192 so no bufferheads are\nattached and the bh walk never sets bdev.  We then submit the bio with a\nNULL block device and crash.\n\nTherefore, truncate the page cache after flushing but before updating\ni_blksize.  However, that's not enough -- we also need to lock out file\nIO and page faults during the update.  Take both the i_rwsem and the\ninvalidate_lock in exclusive mode for invalidations, and in shared mode\nfor read/write operations.\n\nI don't know if this is the correct fix, but xfs/259 found it.\n\ud83d\udccf Published: 2025-06-18T09:33:49.393Z\n\ud83d\udccf Modified: 2025-06-18T09:33:49.393Z\n\ud83d\udd17 References:\n1. https://git.kernel.org/stable/c/64f505b08e0cfd8163491c8c082d4f47a88e51d4\n2. https://git.kernel.org/stable/c/8c5cf440a378801d313eb58be996fdc81a8878a4\n3. https://git.kernel.org/stable/c/c0e473a0d226479e8e925d5ba93f751d8df628e9", "creation_timestamp": "2025-06-18T10:40:08.000000Z"}]}